Distributed multi-robot graph-based SLAM which makes use of condensed measurements to efficiently compress and transmit relevant map information to other robots.
Related publications:
Multi-Robot SLAM using Condensed Measurements.
Localization and SLAM for robot formations using a priori stochastic maps
Cooperative localization of a robot formation using a previously built feature-based stochastic map of the environment. An extension of the EKF algorithm based on the measurement differencing technique is used to deal with time-correlated measurement errors and to improve the localization consistency.
Distributed localization of a robot formation and updating of an a priori map using a EIF-based approach. Conditional Independence properties are applied to work on submaps with a fixed number of features (thus, bounding the cost of map updating), and to make the algorithm scalable in the number of robots. Additionally, the robots only send information matrices of the features acquired since the last synchronization, reducing this way the communication requirements.
Related publications:
Cooperative Minimum Expected Length Planning for Robot Formations in Stochastic Maps
Distributed Localization and Submapping for Robot Formations using a prior map.
Localization of Probabilistic Robot Formations in SLAM.
Cooperative Navigation, Position Tracking and Path Planning for Robot Formations
Using risk maps (probabilistic projection of map features and observations into a grid) the robot formation is able to find an obstacle-free path of minimum collision risk towards its goal. Previously to the experiment, a global minimum risk plan towards the goal is computed using the risk-based map representation of an a priori feature-based stochastic map of the environment. During the experiment, a common local risk map is built from a set of integrated observations to remove redundant information and to obtain a common understanding of the navigation area. Using this cooperative view, the formation is able to deal with unexpected changes of the environment and to increase the efficiency of the on-line replanning process. The cooperative navigation of the robot formation is achieved by means of a virtual spring and damper system which provides a structure flexible and adaptable to the environment.
Adaptability and replanning of a pentagon-shaped robot formation in a simulation environment.
Real experiment with a triangle-shaped robot formation.
Related publications:
Cooperative Minimum Expected Length Planning for Robot Formations in Stochastic Maps
Cooperative Navigation using environment compliant robot formations.
Position Tracking and Path Planning in Uncertain Maps for Robot Formations.